first image fine, but yolo face targeted face on the left, it got pixelated. next photo without yolo

first image fine, but yolo face targeted face on the left, it got pixelated. next photo without yolo the whole photo got pixelated. this is my 100 epoch krea fine tune. i dont understand

�

🍭🎀 𝒜𝓋𝒶 𝐹𝓇𝒾𝑔𝑔 🎀🍭OP•8/20/25, 4:40 PM

i downloaded newest yolo

Sskyrrr got it will try with A100 SXM then

Furkan Gözükara SECourses•8/20/25, 4:40 PM

let me know

�🍭🎀 𝒜𝓋𝒶 𝐹𝓇𝒾𝑔𝑔 🎀🍭first image fine, but yolo face targeted face on the left, it got pixelated. nex...

Furkan Gözükara SECourses•8/20/25, 4:40 PM

newest yolo from downloader app should work

Furkan Gözükara SECourses•8/20/25, 4:41 PM

restart the app

Furkan Gözükara SECourses•8/20/25, 4:41 PM

delete older yolo and redownload to be sure

�

🍭🎀 𝒜𝓋𝒶 𝐹𝓇𝒾𝑔𝑔 🎀🍭OP•8/20/25, 5:00 PM

Ok thanks. But how does that explain the whole image being only pixelated

�🍭🎀 𝒜𝓋𝒶 𝐹𝓇𝒾𝑔𝑔 🎀🍭Ok thanks. But how does that explain the whole image being only pixelated

Furkan Gözükara SECourses•8/20/25, 5:06 PM

it is swarmui error

Furkan Gözükara SECourses•8/20/25, 5:06 PM

happened to me

Furkan Gözükara SECourses•8/20/25, 5:06 PM

or comfyui :d

Furkan Gözükara SECourses•8/20/25, 5:07 PM

when yolo breaks it happens

�

🍭🎀 𝒜𝓋𝒶 𝐹𝓇𝒾𝑔𝑔 🎀🍭OP•8/20/25, 5:12 PM

Ah ok. Maybe ill just try your 2x upscale instead, should make most faces good without face inpaint

�🍭🎀 𝒜𝓋𝒶 𝐹𝓇𝒾𝑔𝑔 🎀🍭Ah ok. Maybe ill just try your 2x upscale instead, should make most faces good w...

AiInfluence•8/20/25, 5:36 PM

check dm

FFurkan Gözükara SECourses delete older yolo and redownload to be sure

�

🍭🎀 𝒜𝓋𝒶 𝐹𝓇𝒾𝑔𝑔 🎀🍭OP•8/20/25, 11:02 PM

Couldn't make yolo model work. Ended up using <segment:face,0.6,0.5> which did work fine

�🍭🎀 𝒜𝓋𝒶 𝐹𝓇𝒾𝑔𝑔 🎀🍭Couldn't make yolo model work. Ended up using <segment:face,0.6,0.5> which did w...

Furkan Gözükara SECourses•8/20/25, 11:10 PM

you probably didnt delete older model and downloaded new one. but this is also fine

Furkan Gözükara SECourses•8/20/25, 11:10 PM

i reported this issue but swarmui dev has no interest to fix

Furkan Gözükara SECourses•8/20/25, 11:10 PM

he also doesnt accept my fixes too

�

🍭🎀 𝒜𝓋𝒶 𝐹𝓇𝒾𝑔𝑔 🎀🍭OP•8/21/25, 8:44 AM

Thats weird. He posted your 90 minute swarmui tutorial video and seemed happy for your support last year. Would think he would listen to your inputs.

�🍭🎀 𝒜𝓋𝒶 𝐹𝓇𝒾𝑔𝑔 🎀🍭Thats weird. He posted your 90 minute swarmui tutorial video and seemed happy fo...

Furkan Gözükara SECourses•8/21/25, 9:00 AM

yep his transformation is weird

Furkan Gözükara SECourses•8/21/25, 9:00 AM

i may fork repo and add some features and make our installers to use them

FFurkan Gözükara SECourses yep his transformation is weird

Dr. Odin•8/21/25, 2:26 PM

I just finished watching the sdxl lora training using massed computes.. and doing my first training as we speak.. however .. training is 20% complete and i already have 2 loras downloaded.. but they look like nothing in the training images.. should i be concerned... just like in the video and all the presets.. donloaded from your patreon.. currently training is at epoch 44 out of 200.. 2 downloaded and they dont look like the training pics.. my prompt has included the ohwx woman...

DDr. Odin I just finished watching the sdxl lora training using massed computes.. and doin...

Furkan Gözükara SECourses•8/21/25, 2:26 PM

you tested on where? swarmui?_

Furkan Gözükara SECourses•8/21/25, 2:27 PM

normally with 20% it should start looking a like

Furkan Gözükara SECourses•8/21/25, 2:27 PM

are you using our presets?

FFurkan Gözükara SECourses are you using our presets?

Dr. Odin•8/21/25, 2:28 PM

yes i am.. downloaded from patreon.. via massed computes i logged into my mimic pc account and tested them in my comfy ui workflow..

DDr. Odin yes i am.. downloaded from patreon.. via massed computes i logged into my mimic ...

Furkan Gözükara SECourses•8/21/25, 2:29 PM

can you also test it inside swarmui installed in massed compute

Furkan Gözükara SECourses•8/21/25, 2:29 PM

i wonder if your training images setup is wrong

Furkan Gözükara SECourses•8/21/25, 2:30 PM

@Dr. Odin we dont allow nsfw here

Furkan Gözükara SECourses•8/21/25, 2:30 PM

you can test example dataset

FFurkan Gözükara SECourses you can test example dataset

Dr. Odin•8/21/25, 2:31 PM

sorry about that.. my apologies

DDr. Odin sorry about that.. my apologies

Furkan Gözükara SECourses•8/21/25, 2:31 PM

example here : https://www.patreon.com/posts/114972274

Patreon

Perfect Quality Example Model Training Images Dataset - Can Be Used...

Get more from SECourses: FLUX, Tutorials, Guides, Resources, Training, Scripts on Patreon

FFurkan Gözükara SECourses you can test example dataset

Dr. Odin•8/21/25, 2:31 PM

okay

Furkan Gözükara SECourses•8/21/25, 2:31 PM

that is a good one to verify your training setup

FFurkan Gözükara SECourses that is a good one to verify your training setup

Dr. Odin•8/21/25, 2:43 PM

lemme see

Kawix•8/21/25, 5:27 PM

How many epochs would you recommend for training 50 images for a 1024x1024 close-up face?

Kawix•8/21/25, 5:34 PM

Small question: if I train my fine-tuned flux model, could I use it for inpainting, or do I need other parameters in koyas_ss to make it work a bit like flux fill?

Kawix•8/21/25, 5:39 PM

Should I do my training with flux fill instead of flux dev?

KKawix Small question: if I train my fine-tuned flux model, could I use it for inpainti...

skyrrr•8/21/25, 5:57 PM

You can use your finetuned checkpoint to do impainting
Finetuning of flux fill isn't possible

KKawix How many epochs would you recommend for training 50 images for a 1024x1024 close...

skyrrr•8/21/25, 5:58 PM

Between 100 and 200. Around 150 is often the sweet spot with flux.dev, it depends on your dataset. I would recommend to save every 20 or 25 epochs, and compare on swarmui using the grid generation tool

Sskyrrr Between 100 and 200. Around 150 is often the sweet spot with flux.dev, it depend...

Kawix•8/21/25, 6:20 PM

ok thx i will do this and follow the tutorial

Kawix•8/21/25, 6:23 PM

@skyrrr By any chance, are you familiar with face detection and Comfy UI? I'm struggling to create a workflow that would detect faces with Yolofacev8 or RetinaFace and then inpaint the face area, which would allow me to regenerate that part with my trained fluxfinetuned. But I can't find a good node. I tried installing “ComfyUI Yolov8” but I got an import error, so I guess it's not compatible. There's no information about this on GitHub.

�

🍭🎀 𝒜𝓋𝒶 𝐹𝓇𝒾𝑔𝑔 🎀🍭OP•8/21/25, 6:28 PM

i'm scared to use this part of kohya, dont wanna waste a night of training and the model files werent saved anywhere...